OK, I've compared graphs
B. XA252 waaaaay down in THD at all power levels, so no real comparison, except same THD Spectra nature (that is more "except")
then, THD levels comparison between B. XA252 SET and B. XA252 SIT clearly shows that SET is uniformly approx. 6db down at each tested power level
in my book, that's pretty good emulation - achieving same behavior and consistency, while still making some desired difference ...... even ignoring that we are emulating nearly 100$ part (plus not so easy to obtain) with 25$ part
edit: for SET - did I said that even famous ∀ Class DEF current distribution is clearly evident - Schade SIT emulation is working as proper SIT in DEF - I like to see it as SIT being error corrector for Mosfet